In [2]:
import os
import pandas as pd
import tensorflow
from keras.preprocessing import image
import numpy as np
import matplotlib.pyplot as plt
import cv2 
from pathlib import Path
from keras.models import Sequential
from keras.layers import Dropout
from keras.layers import Conv2D
In [13]:
def draw_image(img):
    plt.imshow(img,cmap='gray')
    plt.show()
    
p = Path("Desktop/SRIB/img")
images = []
masks = []

cnt = 0
for i in sorted(p.glob("*.jpg")):
    string = str(i)
    img = cv2.imread(string)
    img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    img = cv2.resize(img,(256,256))
    images.append(img)
    cnt += 1
print(cnt)

cnt1 = 0
for i in sorted(p.glob("*.png")):
    string = str(i)
    img = cv2.imread(string)
    img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    img = cv2.resize(img,(244,244))
    masks.append(img)
    cnt1 += 1
print(cnt1)

p = Path("Desktop/SRIB/ShadowImages")
cnt = 0
for i in sorted(p.glob("*.jpg")):
    string = str(i)
    img = cv2.imread(string)
    img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    img = cv2.resize(img,(256,256))
    images.append(img)
    cnt += 1
print(cnt)

p = Path("Desktop/SRIB/ShadowMasks")
cnt1 = 0
for i in sorted(p.glob("*.png")):
    string = str(i)
    img = cv2.imread(string)
    img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    img = cv2.resize(img,(244,244))
    masks.append(img)
    cnt1 += 1
print(cnt1)

X = np.array(images)/255.0
Y = np.array(masks)/244.0
X_ = np.expand_dims(X, axis=3)
Y_ = np.expand_dims(Y, axis=3)
print(X.shape)
print(Y.shape)
print(X_.shape)
print(Y_.shape)
draw_image(X[0])
draw_image(Y[0])
draw_image(X[445])
draw_image(Y[445])

355
355
4085
4085
(4440, 256, 256)
(4440, 244, 244)
(4440, 256, 256, 1)
(4440, 244, 244, 1)
In [11]:
model = Sequential()
model.add(Conv2D(32, (5,5), activation='relu', input_shape=(256, 256,1)))
model.add(Conv2D(32, (5,5), activation='relu'))
model.add(Dropout(0.2))
model.add(Conv2D(1, (5,5)))
model.compile(optimizer='sgd', loss='binary_crossentropy')
model.fit(X_, Y_, epochs=5)

Epoch 1/5
4440/4440 [==============================] - 43s 10ms/step - loss: 0.4620
Epoch 2/5
4440/4440 [==============================] - 41s 9ms/step - loss: 0.3623
Epoch 3/5
4440/4440 [==============================] - 41s 9ms/step - loss: 0.3448
Epoch 4/5
4440/4440 [==============================] - 41s 9ms/step - loss: 0.3334
Epoch 5/5
4440/4440 [==============================] - 41s 9ms/step - loss: 0.3157
Out[11]:
<keras.callbacks.History at 0x1e7d09cc358>
In [12]:
test = []
original = []
p1 = Path("Desktop/img")
cnt = 0
for i in sorted(p1.glob("*.jpg")):
    if cnt > 20:
        break
    string = str(i)
    img = cv2.imread(string)
    img = cv2.resize(img,(256,256))
    original.append(img)
    img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    test.append(img)
    cnt += 1

X_test = np.array(test)/255.0
X_test = np.expand_dims(X_test, axis=3)
print(X_test.shape)
Y_test = model.predict(X_test)*244
Y_test = np.squeeze(Y_test, axis=3)
print(Y_test.shape)
print(type(Y_test))
for i in range(21):
    draw_image(original[i])
    draw_image(Y_test[i])

(21, 256, 256, 1)
(21, 244, 244)
<class 'numpy.ndarray'>